#	Pagina
attuale pagina	/open-h2020/projects/204102/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 2 - MagnonCircuits (Nano-Scale Magnonic Circuits for Novel Computing Systems)

Teaser

Summary

A disturbance in the local magnetic order of a solid body can propagate across a material just like a wave. This wave is named spin wave, and its quanta are known as magnons. Recently, physicists proposed the usage of magnons to carry and process information instead of electrons which are used in electronics. This technology opens access to a new generation of computers in which data are processed without motion of any real particles like electrons. This leads to a sizable decrease in the accompanying heating losses and, consequently, to lower energy consumption, which is crucial due to the ever increasing demand for computing devices. Moreover, unique properties of spin waves allow for the utilisation of unconventional computing concepts, giving the vision of a significantly faster and more powerful next-generation of information processing systems.

The strategic goal of the MagnonCircuits ERC Starting Grant project is to make a transformative change in the data processing paradigm from traditional electronics to magnon-based circuits. The research field of magnonics, which addresses magnon-based data processing, is at the initial stage of its evolution and, to this point, it is located within the physics rather than within the engineering domain. Thus, physical concepts as well as a technological basis should be established on the current stage to achieve the ambitioned goals of all-magnon data processing in the future. Several significant challenges are addressed in the MagnonCircuits project:

â€¢ Magnon structures should be scaled down to lateral sizes below 100 nm and become comparable to the unit sizes in modern electronics. Scalability, which is one of the main advantages promised by the use of spin waves, should be proven.

â€¢ The spin wave properties and dispersion characteristics in such nanostructures should be understood.

â€¢ A methodology should be developed to operate with short wavelengths rather than the nowadays commonly used long wavelengths of dipolar spin waves. Exchange-interaction dominated spin wave allows for the scalability of the magnonic devices and they have significantly higher velocities in nanostructures, ensuring higher data rates and operational speed.

â€¢ Modern spintronics phenomena like spin-orbit torques are very efficient at the nanoscale and should be exploited in order to excite, amplify and detect spin waves.

â€¢ Two-dimensional magnonic structures should be developed on the nanoscale and the spin wave physics within them needs to be understood in order to guide and control spin waves in future magnonic circuits.

â€¢ Nonlinearity is a primary requirement for any computing system. Spin waves are naturally strongly nonlinear and this can be exploited conveniently for data processing. However, scaling down of the structures below 100 nm drastically modifies the nonlinear spin wave physics and the consequent phenomena need to be explored.

â€¢ Finally, new approaches and algorithms for efficient magnon-based computing concepts should be proposed and tested.

Work performed

The project is running successfully and close to the proposed plan. The results were published in a large number of articles and they were advertised on a variety of conferences. The main scientific achievements after the first half of the project are listed below:

â€¢ Yttrium Iron Garnet (YIG) thin films grown by our collaborator C. Dubs (Innovent e.V., Jena) were structured in Kaiserslautern and the corresponding patterning technology was developed. The smallest lateral size of the patterned YIG nanostructures is 30 nm and is well below the planned 100 nm. The first results are reported in [9].

â€¢ For nanostructure characterisation, a new time- and space-resolved Brillouin Light Scattering (BLS) spectroscopy setup was constructed as foreseen in MagnonCircuitsâ€™ work plan and budget. Two different methodologies were implemented: (1) Measurement of the magnetisation precession lifetime and (2) measurement of the spin wave (SW) propagation length in nanostructures. It was demonstrated that the YIG nanostructures preserve the high YIG quality to a large extent.

â€¢ BLS measurements were performed on a Bi-substituted rare earth iron garnet to estimate the exchange stiffness constant [8].

â€¢ Spin-wave properties in nanostructures were studied experimentally and numerically with a particular focus on the nature of the spin wave modes in these confined systems. In cooperation with the external collaborator R. Verba (Institute of Magnetism, Ukraine) a quasi-analytic theory was developed and experimentally confirmed. We have found that the known phenomenon of dipolar pinning disappears if the lateral sizes of the YIG structures go below a critical width around 200 nm which changes the SW mode profiles and their dispersions [9].

â€¢ The generation of coherent SWs by the Spin Hall Effect (SHE) and Spin Transfer Torque (STT) in YIG/Pt nanostructures was evidenced using BLS spectroscopy. Furthermore, the temporal evolution of the auto-oscillations was investigated [1].

â€¢ The fundamentally new phenomenon of Bose-Einstein Condensation by rapid cooling was discovered in YIG/Pt nanostructures [10]. This phenomenon paves the way for the usage of macroscopic quantum magnon states in conventional spintronics.

â€¢ A nanoscaled SW directional coupler was successfully investigated numerically [6] and experimentally (publication in preparation). As opposed to the originally-planned two-dimensional structures, the directional coupler exhibits practically no parasitic reflections and can serve as a SW power splitter, combiner, frequency separator, multiplexer, and (reconfigurable) interconnect.

â€¢ A nanoscale reconfigurable magnonic crystal based on voltage-controlled perpendicular magnetic anisotropy is simulated, pointing a way to low power consuming magnonic applications [3].

â€¢ An array of Abrikosov vortices in a hybrid structure consisting of superconducting and magnetic layers was used to realise a reconfigurable magnonic crystal experimentally [11].

â€¢ Chiral protection of dipole-exchange spin-waves was proposed and studied numerically to minimise parasitic reflections in magnonic circuits [12].

â€¢ A first spin wave majority gate is demonstrated experimentally on the macro-scale [2].

â€¢ Comprehensive studies of multi-magnon scattering and of the nonlinear decrease in magnetisation on the nanoscale were performed using numerical simulations. The latter was employed to design and demonstrate a magnonic half-adder which has already been presented at a few conferences. This device constitutes the first integrated magnonic circuit combining two logic gates (corresponding publication is in preparation). This breakthrough â€“ a simple magnonic structure substitutes 14 transistors in electronics - goes far beyond the originally-proposed ERC research plan.

â€¢ An analogue magnon adder was proposed and studied numerically for future magnon neuromorphic computing operations [7].

â€¢ Two review papers [4, 5] were published

Final results

The successful fabrication of nanostructures from ultrathin YIG and the study of spin wave dynamics within them generally go beyond the state of the art of the research field of magnonics prior to the project. In particular, we would like to underline the successful fabrication and characterization of sub-100 nm large structures of Yttrium Iron Garnet and the understanding of the spin wave properties within them [arXiv:1807.01358] â€“ until the end of the project we foresee a full characterization of the spin wave dynamics in structures with a width around 50 nm and a prove of exchange spin wave propagation within them. In addition, we would like to mention the numerical studies of a nanosized spin wave directional coupler [6] as well as the fabrication of the corresponding nanostructures, which had also not been investigated before in literature. Until the end of the project we expect its full experimental characterization and this way, a first prove of such functional devices on the nanoscale. We would also like to point out that the successful simulation of a magnonic half-adder constitutes the first integrated magnonic circuit and we intend to explore this direction further until the end of the project, aiming at its experimental realization. The further experimental investigations of nonlinear spin wave phenomena at the nanoscale plays a crucial role in this aspect. Ultimately, it should be pointed out that the discovery of a new way to form a Bose-Einstein Condensate of magnons by rapid cooling [arXiv:1612.07305] and the first experimental observation of the interaction between Abrikosov vortices and spin waves [arXiv:1901.06156] are also groundbreakingly new concepts.